loading page

Operational aerosol retrieval at subkilometer resolution using OceanSat-2 OCM over land: SAR algorithm, uncertainties, validation & inter-sensor comparison.
  • Manoj Kumar Mishra,
  • Arundhati Misra,
  • Raj Kumar
Manoj Kumar Mishra
ISRO

Corresponding Author:[email protected]

Author Profile
Arundhati Misra
SAC, Indian Space Research Organization
Author Profile
Raj Kumar
SAC, Indian Space Research Organisation.
Author Profile

Abstract

The OCM sensor onboard OceanSat-2 is providing data in visible and NIR bands. Due to the limited spectral coverage of OCM, the widely used dark-target and deep-blue aerosol algorithms cannot be adapted. Here, a new aerosol optical depth (AOD) retrieval algorithm for OCM (or similar sensors) over land, termed SAC Aerosol Retrieval (SAR) is described. It utilizes two blue bands for the AOD inversion and NIR band to characterize the surface in visible bands without assuming red and NIR bands transparent to aerosols. Unlike the dark-target algorithm, the SAR algorithm can retrieve AOD over bright arid and urban areas too. The uncertainty analysis of SAR suggests a theoretically expected error (EE) envelope of ±(0.06+0.26×AOD) for typical retrieval conditions. OCM AOD over land is retrieved operationally for the first time over Indian and neighboring countries’ landmass at the finest spatial resolution of 0.0070. The SAR algorithm is validated against in-situ AOD measurements during the years 2016-2018 at 21 AERONET stations located in south-Asia. Overall validation using 1900 match-up points shows correlations exceeding 0.8 with 74% of retrievals within EE. The retrievals over cropland, grassland, and mixed land cover types show high (low) correlation (bias), while over bright urban areas somewhat low (high) correlation (bias) is observed. Excluding monsoon season, OCM AOD retrievals show good performance over the year. The performance of MODIS dark-target and OCM AOD, against common in-situ, is close to each other. The study shows that OCM AOD can be used for air quality monitoring/modeling at high spatial resolution.